分整模型在商品价格预测中的应用

首先介绍了时间序列分析中的一个新领域－长记忆分整模型（ＡＲＦＩＭＡ），分析了该模型与传统时间序列模型相比较所体现出的优越性，及其参数估计和预测方法。本文所给出的分整模型不仅反映了传统时间序列模型所不能反映的时间序列长记忆性，而且解决了利用传统方法预测商品价格中的过度参数化问题，从而显著提高了商品价格预测的可靠性。文章还给出了实际案例分析。

The application of ARFIMA model in the price forecasting of commodity

To begin with the introduction of ARFIMA model, a new branch of Time Series Analysis,the paper analyzes the advantage of ARFIMA model over the traditional time series models and gives the parameter estimation method and the forecasting formula of ARFIMA model.The ARFIMA model not only reflects the long memory property of time series which cannot be reflected by traditional time series models,but also solves the problem of over parameterization which occurs when applying traditional time series models to the price forecasting of commodity.Finally,a positive study is presented.